Carrier and Noise Measurements in the Medium Wave Band for Urban Indoor Reception of Digital Radio

BTS-11-111 1 Carrier and Noise Measurements in the Medium Wave Band for Urban Indoor Reception of Digital Radio Igor Fernandez, member IEEE, Pablo Angueira, member IEEE, David De la Vega, member IEEE, Iván Peña, member IEEE, David Guerra, member IEEE, Unai Gil, member IEEE Abstract This paper presents a study of indoor digital radio reception in the Medium Wave band. The results are based on a large database of measurements in indoor locations in cities in northern Spain in frequencies from 600 to 1600 khz. The paper proposes a prediction model for indoor field strength. The results are combined with indoor noise levels to study the statistics of available C/N (Carrier to Noise) ratios. The model is intended to help broadcasters planning digital broadcasting networks in Medium Wave for indoor reception. Signal time and space variability have also been studied and reference values are provided in the paper. The paper includes curves to calculate indoor losses in respect to ground wave propagation values provided by ITU-R P.368. Index Terms Indoor, Medium Wave, Digital Radio Mondiale. HD-Radio, IBOC P I. INTRODUCTION ORTABLE reception is one of the major advantages of audio broadcasting in general, and also in the MW (Medium Wave). Traditionally, a significant amount of MW listeners accessed the services using a portable radio located indoors. During the last two decades, the market for MW broadcasting has been declining in most countries, both in number of services, receivers and listeners. The irruption of Digital Radio standards in the late nineties has revitalized the MW band. These standards are DRM (Digital Radio Mondiale) [1]-[5] and HD-Radio [6]-[8]. Indoor reception is a challenge for many radiocommunication services but researches have only been carried out in higher frequency bands [9]-[1]. Even in the most favorable cases, planning for indoor reception requires careful frequency, site and power selection. Building penetration losses, spatial variability and high man-madenoise values should also be taken into account in order to plan indoor coverage in MW. Manuscript received June 3, 011. This work was supported in part by the University of the Basque Country and Basque Government. All authors are with the Electronics and Telecommunications Department, University of the Basque Country, Bilbao 48013 Spain (phone: 0034-946013918; fax: 0034 94601459; e-mail: igor.fernandez@ehu.es). The signal and noise behavior in MW has been studied for outdoor reception in tests over the last five years []-[4]. Indoor noise levels in MW have been recently characterized in [5] observing that there has been a remarkable increase in the average and peak noise values, but the required information for indoor signal reception and C/N (Carrier to Noise) ratio have not been obtained yet. Some preliminary indoor signal and noise results in MW were presented in [6] but with a reduced indoor location database. This paper studies indoor reception on the basis of an extensive measurement campaign in different buildings located in urban environments of northern Spain. The work was designed to include an analysis of spatial and time variability of both signal and C/N values, with the final objective of providing a tool to obtain an estimation of the indoor coverage in urban environments. The study has been performed over a representative number of frequencies in the MW band ranging from 600 to 1600 khz. The signal reception behavior is compared to the results obtained in [5] for indoor noise levels in MW to provide an estimation of the available C/N ratios within the service area of a MW Broadcast transmitter in an urban environment. The paper is organized as follows. Section II contains all the relevant information related to the field tests, including broadcast networks and reception system. The measurements methodology is also explained in this section. Section III presents the results of the field test and the analysis of the data. First, a C/N analysis is carried out for present analogue broadcasting systems in MW. Then, indoor signal spatial and time variability is statistically analyzed in order to provide a model and some planning parameters. Finally, prediction curves for median signal, noise and C/N are provided together with their probability density distributions. The paper finishes with a summary of the most relevant conclusions in Section IV. II. MEASUREMENTS CAMPAIGN A. Transmission Systems and Measurements in Buildings The field tests have been based on measurements done on existing AM (Amplitude Modulation) commercial transmissions in the cities of Bilbao and Vitoria-Gasteiz in 009 and 010. Nevertheless, the results should be directly

BTS-11-111 applicable to the digital systems (DRM, HD-Radio or another future system on this band) since AWGN (Additive White Gaussian Noise) channels have been considered for ground wave propagation in MW at daytime without time and frequency dispersion []. As shown in Table I six different MW frequencies have been selected in order to have samples on the low, middle and upper sides of the band. TABLE I ANALYZED FREQUENCIES Frequency Network Antenna (khz) Operator height (m) Power (kw) 61 RNE 90 10 639 RNE 90 40 990 SER 60 5 1197 EiTB 8.5 10 1305 RNE 60 6 160 EiTB 8.5 5 Fig. 1. Transmitters and buildings in Bilbao. Measurements in five office buildings (OB1 to OB5) and ten household buildings (HB1 to HB10) in Bilbao have been taken. The measured carriers in those buildings were 639, 990 and 1305 khz. A set of another six buildings have been selected in a different city (Vitoria-Gasteiz) where carriers at 61, 1197 and 160 khz were measured in an office building (OV1) and five household buildings (HV1 to HV5). Fig. 1 and Fig. show buildings where measurements have been taken and transmitters in both cities respectively. Table II summarizes the number of locations where four minutes measurements were taken for each carrier frequency. TABLE II NUMBER OF LOCATIONS AND BUILDINGS Fig.. Transmitters and buildings in Vitoria-Gasteiz. Area Bilbao Vitoria- Gasteiz Frequencies (khz) 639, 990, 1305 61, 1197, 160 Building Building type identifier Locations OB1 40 OB 45 Office OB3 7 OB4 33 OB5 15 HB1 1 HB 1 HB3 7 HB4 9 Household HB5 7 HB6 10 HB7 1 HB8 5 HB9 8 HB10 7 Office OV1 41 HV1 16 HV 16 Household HV3 1 HV4 9 HV5 8 The measurement points were planned previously to follow a uniform spatial distribution across each floor in office buildings, where access to instances was easier. In home buildings, almost every indoor location was sited in hallways of every floor, since access to every particular home was not feasible. Only in household buildings HV and HV3 access to more instances was possible so that a uniform location spatial distribution was possible for these buildings too. Table III summarizes the most relevant characteristics of buildings where measurements were taken. TABLE III MAIN CHARACTERISTICS OF BUILDINGS Building type Material Floors Environment Windows Office Concrete 3-5 Urban Large Household Concrete 6-11 Urban Small

BTS-11-111 3 The total number of locations was 360 and in each location measurements were taken along four minutes for each carrier frequency. In order to complete the analysis, at 16 locations inside buildings in Bilbao, measurements were taken along approximately 48 hours and data files were cut in one daytime hour files so that a total number of 9 files of one hour were stored for each frequency to analyze the time variability in a time period longer than 4 minutes. The same procedure was carried out in locations in Vitoria for a total number of 33 daytime hours for each frequency. B. Methodology The measurement system was installed on a portable trolley (a block diagram is presented in Fig. 3). A calibrated passive antenna ETS Lindgren 3303 was mounted on the trolley at 1.5 m above floor level and the field strength measurements were done using the R&S ESPI3 meter. This equipment was remotely controlled using software specifically designed for the test. The measurement methodology consisted of recording four minutes of continuous signal data, for each frequency, as explained in the measurement campaign intended for noise and published in [5]. ETS Lindgren passive antenna R&S ESPI3 field meter Control laptop variability and short term time variability. Since only few long term measurements were taken, they will only be used to analyze the long term time variability in order to make comparisons with the short term time variability. III. RESULTS A. Indoor (C/N) Values Fig. 4 shows the histogram of measured indoor C/N values achieved with the present broadcasting planning used in cities of Bilbao, Vitoria-Gasteiz and surroundings for the analogue AM system. It should be noted that the C/N value in each location has been calculated subtracting the median noise field strength value measured in [5] in adjacent free channels to the median signal carrier field strength value measured in the campaign described in this paper, since locations where measurements were taken for signal carriers and noise were the same in both campaigns. 675, 1035 and 14 khz free adjacent channels from [5] have been selected for locations in Bilbao and 630, 14 and 160 khz free adjacent channels for locations in Vitoria. Minimum required C/N values for quasi error free reception vary from around 8 db to 18 db in DRM for robustness mode A (typical for ground wave propagation) and 9 khz channeling in MW [], [5]. In the other hand, for HD-Radio, the minimum required C/N value could be set in approximately 11 db for the hybrid 30 khz mode and in 13 db for the all digital 0 khz mode in MW [8]. As observed in Fig. 4, the C/N values would be under the minimum required C/N value in a significant number of indoor locations. 90,0 80,0 Fig. 3. Portable measurements system block diagram. It should be noted that the measurements were carried out during daytime so no long distance ionospheric interference was expected. Field meters were configured with 100 Hz resolution bandwidth for AM signal carrier measurements in order to measure only the carrier level and skip any dependency on modulation, whereas for noise measurements carried out in [5], typical 9 khz channeling was maintained. Signal values were recorded using a RMS detector with an integration time of 400 ms (duration of a DRM frame). This integration period has been used in previous DRM trials and was kept for future comparison purposes []-[4]. Field strength value for each location was calculated adding the antenna K factor specifically calibrated and provided by the manufacturer of the antenna for each frequency to the signal voltage level indicated in the R&S ESPI3 meter. Finally, it should be noted that four minutes measurements will be used to analyze the modulated signal s spatial Number of occurrences 70,0 60,0 50,0 40,0 30,0 0,0 10,0 0,0-3,5-0,7-17,9-15,1-1, -9,4-6,6-3,8-1,0 1,8 4,6 7,4 10,3 13,1 15,9 18,7 1,5 4,3 7,1 9,9 3,8 35,6 38,4 41, 44,0 46,8 49,6 5,4 55,3 58,1 60,9 greater C/N (db) Fig. 4. C/N values histogram in indoor locations in Bilbao and Vitoria- Gasteiz. B. Indoor Signal Variability All the median signal carrier field strength samples were normalized for 1 kw transmitted power at every frequency for the rest of the analysis carried out in this paper, in order to be able to compare different frequencies altogether. Fig. 5 shows normalized median field strength values at each location for

BTS-11-111 4 990 and 1305 khz, respectively (transmitter sites are relatively close to each other in Bilbao) together with logarithmic tendency lines with distance. Results show that indoor field strength decreases with distance to transmitter and also with frequency. The high variability with locations in almost every building, but especially in buildings where more measurements have been taken is also remarkable. In fact, differences up to 0 db in received levels have been found in locations distanced few meters to each other, so the spatial variability is very high as it happened in other indoor campaigns in higher frequency bands [9], [15], [16], [1]. 10,00 110,00 100,00 90,00 80,00 70,00 60,00 50,00 40,00 3,40 3,50 3,60 3,70 3,80 3,90 4,00 4,10 4,0 4,30 log 10 d(m) 990 khz 1305 khz Fig. 5. Indoor normalized median field strength level in each location at 990 and 1305 khz. It has not been observed significant difference in behavior of signal spatial variability in offices and households. Similar conclusions are obtained for 1197 and 160 khz in Vitoria-Gasteiz, as it can be seen in Fig. 6. 110,00 100,00 90,00 80,00 70,00 60,00 50,00 40,00 30,00 3,80 3,85 3,90 3,95 4,00 log 10 d(m) 1197 khz 160 khz Fig. 6. Indoor normalized median field strength level in each location at 1197 and 160 khz. Table IV and Table V show standard deviations with locations in each building in Vitoria-Gasteiz and Bilbao respectively. It is seen that in some buildings (especially where more measurements have been taken), high standard deviations with locations are obtained at all frequencies. TABLE IV STANDARD DEVIATIONS WITH LOCATIONS IN EACH BUILDING IN BILBAO Building σ 639 khz (db) σ 990 khz (db) σ 1305 khz (db) OB1 11. 10.3 1.3 OB 10.4 13. 9.5 OB3 13.6 10.7 11.4 OB4 9.7 8.7 9.3 OB5 7.0 10.3 1.4 HB1 1.3 5. 10.9 HB 1.5 3..8 HB3 7.9 11.7 3.7 HB4.5 4. 6.4 HB5 3.8 6.7 14.3 HB6.5 6.8 6.8 HB7 5.8 7.8 4. HB8 1.0.1.0 HB9 4.9 3.9 6.7 HB10 1.5 4.8 6. TABLE V STANDARD DEVIATIONS WITH LOCATIONS IN EACH BUILDING IN VITORIA- GASTEIZ Building σ 61 khz (db) σ 1197 khz (db) σ 160 khz (db) OV1 6.4 10.4 10.5 HV1 5.8 4.5 8. HV 8.7 5. 7.8 HV3 7.6 14.1 13.6 HV4 10.4 9.7 13. HV5 3.1 6.1 6.1 In some buildings, it has been observed that indoor field strength level tends to increase with floor level. As an example, Fig. 7 shows the median field strength levels with locations in each floor of buildings OB1 to OB5 altogether (all them almost at same distance to transmitters) at each analyzed frequency. Standard deviations with locations have also been included as error bars observing again the high spatial variability even in locations at the same floor. 100,0 96,0 9,0 88,0 84,0 80,0 76,0 7,0 68,0 64,0 60,0 56,0 5,0 48,0 44,0 40,0 (0) Floor 0 (1) Floor 1 () Floor (3) Floor 3 (4) Floor 4 (0) (1) () (3) (4) (0) (1) () (3) (4) (0) (1) () (3) (4) 639 990 1305 Frequency (khz) Fig. 7. Median field strength as a function of floor number and frequency in OB1 to OB5 buildings and standard deviations with locations. Finally, Table VI shows the time RMS standard deviation with all locations in the database when observed along a 4 minute period compared to that over a 1 hour slot at each measured location. In every case, very low values are obtained for both 4 minute and 1 hour periods. As a conclusion, results suggest that modulated signal time variability could be ignored in any case.

BTS-11-111 5 TABLE VI TIME RMS STANDARD DEVIATIONS WITH LOCATIONS Freq. (khz) 4 minutes 1 hour 61 0.8 0.4 639 0.8 0.5 990 0.9 0.7 1197 0.7 1.0 1305 0.8 0.5 160 0.6 0.4 C. Planning Parameters and Prediction Curves Once the variability and dependency with several factors has been analyzed, it is now intended to develop a prediction model for indoor median field strength (C) and C/N ratio, using linear regressions and trying to minimize the Mean Square Error (MSE). The following formula predicts the median field strength value in MW indoors (C), as a function of distance to transmitter and frequency for indoor reception. It was obtained from all measurements carried out in Bilbao and Vitoria- Gasteiz by linear regressions: C ( dbµ V / m) = 4.1 19.0 log 10 ( d ( m)) 6. log 10 ( f ( khz )) σ = 1. db error Due to the high variability within locations, the standard deviation of the prediction error is also high (1. db). Adding a variable to compute the effect of floor level, prediction error is not significantly reduced, although it has been seen in some buildings that indoor field strength level increases slightly with floor level. In the other hand, ground wave propagation curves reported in ITU-R P.368 [7] have been usually used to predict the field strength level in rural environments in MW for 1 kw transmitted power. These curves do not take into account penetration losses in buildings in urban environments. In Fig. 8 free-space propagation and ground-wave prediction curves in ITU-R P.368 have been depicted as a function of distance (with conductivity σ=0.005 S/m and relative permittivity ε r =30 in Bilbao and σ=0.016 S/m and ε r =30 in Vitoria [8]) for representative frequencies of 600 and 1600 khz. Predicted indoor curves of C(dBµV/m) in equation (1) have been also included for these two frequencies. 110 100 90 80 70 Indoor 600 khz Indoor 1600 khz ITU-R P.368 σ=16ms/m (600 khz) ITU-R P.368 σ=16ms/m (1600 khz) 60 ITU-R P.368 σ=5ms/m (600 khz) ITU-R P.368 σ=5ms/m (1600 khz) Free Space 50 1 10 Distance (km) Fig. 8. ITU-R P.368 curves and indoor estimation curves at 600 and 1600 khz, σ=0.005 S/m and σ=0.016 S/m, and ε r=30. (1) The functions obtained from the measurements follow the tendency of ITU curves up to around 10 km. But for higher distances, ground wave propagation curves decrease faster with distance than predicted indoor curves so these curves or prediction formula (1) is not valid for long distances. The analyzed buildings are not farther than 0 km from transmitters, so this is why equation (1) fits data for short distances but it is not valid for longer distances. In order to include both dependency with distance and ground properties of ITU-R P.368, losses of indoor field values respect to ground wave curves are proposed now to be analyzed. Using then the ground wave propagation values as a reference (with conductivity σ=0.005 S/m and relative permittivity ε r =30 in Bilbao and σ=0.016 S/m and ε r =30 in Vitoria) and calculating the difference between these values and real values obtained in indoor locations in Bilbao and Vitoria-Gasteiz, the following regression formula is obtained to quantify what here has been called LITU (Losses respect to ITU) as a function of frequency: LITU ( db) = 4.1+ 0.5 log 10 ( f ( khz)) σ = 11. db () error 8 This formula () has been graphically represented in Fig. 9, observing that losses respect to ground wave curves vary from around 15 db at 600 khz to 4 db at 1600 khz: LITU (db) 8,0 6,0 4,0,0 0,0 18,0 16,0 14,0 1,0 10,0 500,0 700,0 900,0 1100,0 1300,0 1500,0 1700,0 Frequency (khz) Fig. 9. Predicted losses respect to ITU-R P.368 curves (LITU). In order to validate LITU prediction equation (), some indoor signal measurements (96) were also taken in 7 buildings in Madrid in 008 [5], [9]. An experimental 1,5 kw DRM transmission was analyzed at 160 khz, and an active loop antenna Wellbrook ALA 1530 was used to enhance digital signal reception with a slightly better sensitivity in the measurement system. It was demonstrated in a PhD thesis [30] that σ=0.016 S/m and ε r =30 fitted correctly with an outdoor extensive measurements campaign carried out in Madrid. Applying LITU equation () for measurements carried out in buildings of Madrid and for approximately 0% of locations in Bilbao and Vitoria-Gasteiz (locations in buildings OB3, OB5, HB, HB5, HB9, HV1 and HV5 were randomly selected), and comparing them with real LITU values, a low mean prediction error of 0. db is obtained so it seems that, on

BTS-11-111 6 average, prediction formula () fits correctly (if we apply equation () just for indoor measurements in Madrid, the prediction error is also low (1. db)). Consequently, it is recommended to predict the indoor median field strength value subtracting the correspondent LITU value to the field strength reported in ITU-R P.368 as follows: C( dbµ V / m) = C ITU R P.368 ( dbµ V / m) + 4.1 0.5 log 10 ( f ( khz)) σ = 11.8dB (3) error This way, indoor predicted curves would follow the tendency of ground wave curves with distance and ground properties but would also compute the losses when penetrating in cities and in buildings. Fig. 10 represents, as an example, the predicted indoor curves at 600 and 1600 khz with σ=0.016 S/m and ε r =30. 110 100 90 80 70 60 50 1 10 Distance (km) Free Space ITU-R P.368 - LITU 600 khz ITU-R P.368 - LITU 1600 khz Fig. 10. Final predicted indoor field strength curves at 600 and 1600 khz (σ=0.016 S/m and ε=30). From noise measurements database employed in [5], the following prediction formula could be obtained for noise field strength (N) as a function of frequency by linear regressions without taking into account the building type (office buildings and households were differentiated in [5]), since prediction error is almost the same: N ( dbµ V / m) = 149.5 8. log 10 ( f ( khz )) σ error = 10. db Finally, prediction formulas for C/N (for 1 kw transmitted power) could be obtained as the difference between prediction formula (3) for signal or carrier C and prediction formula (4) for noise N: C / N( db) = C ( dbµ V / m) 107.4 + 7.7 log10 ( f ( khz)) (5) ITU R P.368 D. Statistical Distributions The possible adjustment of data to theoretical probability distributions has been studied in this section for indoor signal and noise field strengths. Since signal and noise [5] time variability could be ignored, only spatial variability has been analyzed. The candidate distribution functions were Gaussian, (4) Log-Normal, Rice, Rayleigh, Weibull and Nakagami distributions [31]. It has been decided to normalize the median signal field strength values in each location in order to skip the dependency of distance and frequency and, this way, to be able to apply statistical techniques to the whole dataset obtained from this normalization. LITU values have been analyzed instead of signal field strength values in order to skip the dependency of signal strength with distance. Then, in order to skip the dependency with frequency, each LITU value has been normalized with mean LITU value obtained at each frequency in the whole database. Applying the Maximum Likelihood Estimate method (MLE) [3], the highest values for the logarithmic likelihood function have been obtained for the Log-Normal distribution. Furthermore, this statistical distribution is the only one which passes the Chi-square test [3], so that this distribution is the only one which fits correctly with data. This is why the Log- Normal distribution is proposed as the reference to characterize the MW indoor signal field strength s spatial variability with σ=1.359 standard deviation, not of the linear variable but of the neperian logarithm of this variable [31]. Median noise field strength s statistical distribution should be also obtained (as it was not done in [5]), so a similar procedure for noise database in [5] has been applied, normalizing median noise field strength in each location with respect to the mean noise field strength value obtained at each frequency, in order to skip the dependency with frequency in each location (it was observed in [5] that distance had not any influence on received noise field strength level). Applying then MLE and Chi-square tests to the whole normalized dataset, Log-normal distribution is again the only one distribution which passes the Chi-square test and it is the distribution which obtains the highest values for the logarithmic likelihood function, so the Log-normal distribution is proposed again as the reference to characterize the MW indoor noise field strength s spatial variability with σ=1.10 standard deviation. Knowing that indoor signal C and noise N field strength values are lognormally distributed with 1.359 and 1.10 standard deviations respectively, the following random variable would be normal or Gaussian since is a difference between two normal independent variables: ln( C / N ) = ln( C ) ln( N ) (6) It is concluded that C/N is also lognormally distributed or, what is the same, ln(c/n) is normally distributed. So everything we need to characterize ln(c/n) normal random variable is to calculate its mean value and variance or standard deviation [31]: E { ln( C / N )} = E{ ln( C) } E{ ln( N )} (7) { ln ( C / N )} E { ln( C / N )} σ ln( C / N ) = E (8)

BTS-11-111 7 The mean or median value can be calculated directly from C/N prediction formula (5) for the desired distance, frequency and ground properties, adding the following unit change: C / ( db 0 { ln( C / N )} ln(10 N ) / ) E = (9) The variance, however, requires the following calculation: σ ln( C / N ) E = E = σ { ln ( C) + ln ( N ) ln( C) ln( N )} { ln( C) } E { ln( N )} + E{ ln( C) } E{ ln( N )} { ln ( C) } E { ln( C) } + E{ ln ( N )} E { ln( N )} ln( C ) = E + σ ln( N ) = 1.359 So the standard deviation: + 1.10 = 3.101 = = (10) σ ln( C / N ) = σ ln( C) + σ ln( N ) = 1.761 (11) This standard deviation and mean value calculated in equations (5) and (9) should be employed in the Log-normal distribution in order to calculate the C/N exceeded in the desired percentage of locations. IV. CONCLUSIONS Indoor signal field strength s spatial and time variability in MW has been analyzed, observing high standard deviations with locations and very low time standard deviations, so this variability could be ignored. Combining this analysis with recent indoor noise level s studies, it has been seen that the minimum required C/N values for digital standards are not achieved in many locations with the present broadcasting planning for analogue services. In order to facilitate a proper network planning for indoor coverage of these digital standards, prediction formulas for indoor field strength levels and C/N ratios based on MSE regressions of a large database of measurement results have been presented. Indoor signal, noise and C/N values have been statistically characterized and Log-normal distribution is the one which best fits data with σ=1.359, σ=1.10 and σ=1.761 standard deviations, respectively. This way, combining prediction formulas and provided Log-normal standard deviations, C/N values exceeded in desired percentage of indoor locations could be estimated at a given frequency, distance and surrounding ground properties. ACKNOWLEDGMENT This work has been partially funded by the UPV/EHU and the Basque Government. REFERENCES [1] International Telecommunication Union ITU-R Recommendation BS.1514-1, System for digital sound broadcasting in the broadcasting bands below 30 MHz, Oct. 00. [] International Telecommunication Union ITU-R Recommendation BS.1615, Planning parameters for digital sound broadcasting at frequencies below 30 MHz, 003. [3] International Electrotechnical Commission, Digital Radio Mondiale (DRM) Part1: System Specification, IEC 67-1, Mar. 003. [4] European Telecommunication Standard Institute, Digital Radio Mondiale (DRM): System specification, ETSI ES 01 980 v..1, Oct. 005 [5] International Telecommunications Union ITU-R Report BS.144, Planning Parameters and Coverage for Digital Radio Mondiale (DRM) Broadcasting at Frequencies Below 30 MHz, May 009. [6] Federal Communications Commission Digital Audio Broadcasting Systems and Their Impact on the Terrestrial Radio Broadcast Service. First Report and Order, MM Docket No. 99-35; FCC 0-86, 17 FCC Rcd 19990, Oct. 00 [7] Federal Communications Commission, Digital Audio Broadcasting Systems and Their Impact on the Terrestrial Radio Broadcast Service. Final rule, 47 CFR Part 73, MM Docket No. 99-35; FCC 07-33, August 007. [8] National Radio Systems Committee Std. NRSC 5-A, In-Band/On- Channel Digital Radio Broadcasting Standard, Sept. 005. [9] W. Joseph, L. Verloock, D. Plets, E. Tanghe and L. Martens, Characterization of Coverage and Indoor Penetration Loss of DVB-H Signal of Indoor Gap Filler in UHF Band, IEEE Trans. Broadcast., vol. 55, no. 3, pp. 589 597, Sept. 009. [10] D. Plets, W. Joseph, L. Verloock,, L. Martens, Hugo Gauderis and Etienne Deventer, Extensive Penetration Loss Measurements and Models for Different Building Types for DVB-H in the UHF Band, IEEE Trans. Broadcast., vol. 55, no., pp. 13, June 009. [11] A. Martinez, D. Zabala, I. Peña, P. Angueira, M.M. Velez, A. Arrinda, D. De la Vega and J.L. Ordiales, Analysis of the DVB-T Signal Variation for Indoor Portable Reception, IEEE Trans. Broadcast., vol. 55, no. 1, pp. 11 19, March 009. [1] J.T. Ong, H. Yan, S.V.B. Rao and G. Shanmugam, Indoor DTV Reception: Measurement Techniques, IEEE Trans. Broadcast., vol. 50, no., pp. 19 199, June 004. [13] R.E. Wetmore, M. Radford, D. Schnelle, J. Meehan, R.A. Kennedy and D. Miniutti, Joint VSB Terrestrial Digital TV Indoor Reception Project, IEEE Trans. Broadcast., vol. 48, no. 4, pp. 61 73, Dec. 00. [14] R. Schiphorst, N.A. Moseley, A.C. Aarden, M. Heskamp and C.H. Slump, A T-DAB Trial Using a Low-Mast Infrastructure, IEEE Trans. Broadcast., vol. 54, no. 3, pp. 356 370, Sept. 008. [15] Seidel, S. Y., and Rappaport T.S., 914 MHz Path Loss Prediction Models for Indoor Wireless Communications in Multifloored Buildings, IEEE Transactions on Antennas and Propagation, vol. 40, No., pp. 07-17, February 199. [16] A. Davidson, C. Hill, Measurement of Building Penetration into Medium Buildings at 900 and 1500 MHz, IEEE Transactions on Vehicular Technology, vol. 46, No. 1, February 1997. [17] S.Mason, Indoor Reception of DAB Consequences for Planning and Implementation, EBU Technical Review, July 004. [18] J. A. Green, B.A., Building Penetration Loss Measurements for DAB Signals at 11 MHz, BBC Research Department Report, 199. [19] Hoffman H.H. and Cox D.C., Attenuation of 900 MHz radio waves propagating into a metal building, IEEE Trans., AP-30, No. 4, pp. 808-811, 198. [0] R.R. Murray, H.W. Arnold and D.C. Cox, 815 MHz Radio Attenuation Measured within a Commercial Building, in Proc. of the Antennas and Propagation Society International Symposium, 1986. [1] L. P. Rice, Radio Transmission into Buildings at 35 and 150 mc, The Bell System Technical Journal, January 1959. [] D. Guerra, G. Prieto, I. Fernandez, J. M. Matías, and P. Angueira, Medium wave DRM field test results in urban and rural environments, IEEE Trans. Broadcast., vol. 51, no. 4, pp. 431 438, Dec. 005. [3] G. Prieto, M. Vélez, A. Arrinda, U. Gil, D. Guerra and D. De la Vega, External Noise Measurements in the Medium Wave Band, IEEE Trans. Broadcast., vol. 53, no., pp. 553 559, June 007. [4] U. Gil, D. Guerra, P. Angueira, J. Morgade,D. De la Vega and M. Velez, DRM 0 khz Simulcast Field Trials in the Medium Wave Band in

BTS-11-111 8 México D.F., IEEE Trans. Broadcast., vol. 54, no. 1, pp. 53 61, March 008. [5] I. Fernandez, P. Angueira, I. Landa, A. Arrinda, D. Guerra and U. Gil, Indoor Noise Measurements in the Medium Wave Band, Electronics Letters, vol. 46, issue 1, 5 August 010. [6] I. Fernandez, P. Angueira, I. Landa, A. Arrinda, J.L. Ordiales, D. De la Vega and M.M. Velez, Urban Indoor Signal and Noise Measurements in the Medium Wave Band, in Proc. of the Loughborough Antennas and Propagation Conference, November 009. [7] International Telecommunications Union ITU-R Recommendation P.368-9, Ground-Wave Propagation Curves for Frequencies between 10 khz and 30 MHz, 007. [8] International Telecommunications Union ITU-R Recommendation P.83-, World Atlas of Ground Conductivities, 1999 [9] U. Gil, D. Guerra, L. Del Amo, J. Masdeu, DRM field trials for urban coverage planning in Spain, European Broadcasting Union technical review, 008. [30] D. Guerra, PhD Thesis Estudio de parámetros para la planificación de redes DRM en la banda de Onda Media diurna, Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco (UPV/EHU), Sept. 006. [31] International Telecommunications Union ITU-R Recommendation P.1057-, Probability Distributions Relevant to Radiowave Propagation Modelling, 007. [3] D. Guerra, U. Gil, D. De la Vega, G. Prieto, A. Arrinda, J.L. Ordiales and P. Angueira, Medium Wave Digital Radio Mondiale (DRM) Field Strength Time Variation in Different Reception Environments, IEEE Trans. Broadcast., vol. 5, no. 4, pp. 483 491, Dec. 006.